Arrakis.py map script discussion

[Deliverator]

I think that only shows that spice is important enough to the game right now. You can get high commerce by just spamming cottages, and other means.

 

[davidlallen]

Are you sure this isn't just a matter of perception?

I completely agree. One solution is to accept that different maps play quite differently. The only drawback to this is any tuning change we make needs to be tested on both maps, and the results from one map may contradict the other. Another solution is to remove dunipelago, which has been voted down.

Another possible solution is to make smaller islands on the arrakis script. I am sure if there was already a python variable for that, you would have mentioned it. So would that be an easy user-requested enhancement for you to add?

 

[Ahriman]

Are you sure this isn't just a matter of perception? I've only played on Arrakis, and I haven't noticed any problems, maybe because I just don't know any better. All the civs seem to progress just fine.

Yes. Just because the civs progress doesn't mean that:
a) spice constitutes a significant portion of the economy. Spice feels unimportant on Arrakis script.
b) arrakis spice vs arrakis paradise are balanced. If paradise vs spice are balanced on duneipegalo, then Paradise will be a no-brainer on Arrakis. If they're balanced on Arrakis, then spice economy will be a no-brainer on Duneipegalo.

I think that only shows that spice is important enough to the game right now. You can get high commerce by just spamming cottages, and other means.

Are you missing a not between "is" and "important"?

Another possible solution is to make smaller islands on the arrakis script.

And a third is to add some smal (~10-12 tile)l outlying islands outside the main ring.

 

[davidlallen]

And a third is to add some smal (~10-12 tile)l outlying islands outside the main ring.

Is it worthwhile to try adding *tiny* islands, 2-3 tiles, in that area? I could probably do that with some random numbers. Find a spot, which has no land for 2 tiles in each direction, and add a few random land squares right there.

 

[Ahriman]

Is it worthwhile to try adding *tiny* islands, 2-3 tiles, in that area?

Not really. They wouldn't really be worth putting a city on, and the human player would exploit them to build forts to get cultural area to get some extra spice, which the AI woudn't understand how to do.

The islands need to be large enough to build a functional city on, which means 2-3 water sources at minimum.

 

[Deliverator]

Are you missing a not between "is" and "important"?

Yes, I meant to say, 'I think that only shows that spice is not important enough to the game right now.'

 

[Ahriman]

So we're 1-all on saying the opposite of what we mean

 

[Deliverator]

Indeed.

 

[davidlallen]

Not really. They wouldn't really be worth putting a city on, and the human player would exploit them to build forts to get cultural area to get some extra spice, which the AI woudn't understand how to do.

Suppose we had tiny islands, which humans could build forts upon, and suppose we somehow gave the AI enough interest in building those forts also. I'm not sure how, but perhaps a clever AI coding person can figure that out. It doesn't seem to fit ""reality"" to have big islands with cities all the way out in the deep desert. I could accept rock outcroppings big enough to build a fort.

 

[Deliverator]

We have this modcomp already in the mod.
http://forums.civfanatics.com/showthread.php?t=283686

How about giving harvesters a 1 plot cultural radius or even a BFC? That would enable you to work your way out into the desert. Currently the tags are set to these for harvester:

<bSpreadCultureControl>0</bSpreadCultureControl>
<iCultureBorderRange>1</iCultureBorderRange>

I'm not sure the reason for this, but I assume the culture border has no effect because the first setting is 0.

 

[Ahriman]

It doesn't seem to fit ""reality"" to have big islands with cities all the way out in the deep desert

How is it "real" to only have a single ring of rock around the pole that cities can be built on?

That seems very unreal to me; why is it that high altitude rock would happen to occur around the pole, but not anywhere else? That's not how geology works. You'll have the raised rock areas along plate boundaries, not in just a single circle around the pole.

Having offshore islands of rock a few tiles away from the main ring seems fine. Basically I would think of it this way; its really part of the same continent, but its elevations was lower and filled up with sand.
Like the british isles sitting off of Europe; if the sea level were a little lower they'd still be part of Europe, but with the higher sea level the low bit in between them got filled with water.
This is the kind of thing I'm imagining.

How about giving harvesters a 1 plot cultural radius or even a BFC? That would enable you to work your way out into the desert.

I'm not sure this is a great solution, because of how random it is. One player might be able to work their culture out over a huge area, because of successive spice blooms that just happened to match up next to each other, while another player might not be able to work out at all, because the next spice bloom happened just 1 tile further away.

And what happens when the harvester disappears? You get left with weird floating blobs of culture.

 

[davidlallen]

How is it "real" to only have a single ring of rock around the pole that cities can be built on?

Sorry, I used quadruple quotes without defining it. I was referring to reality according to the Dune books, not necessarily real geography. In the books, the description of the planet is that the only usable land for creating cities is in the polar regions. Granted, the archipelago mapscript does not do this. But now that the arrakis mapscript exists, I would prefer not to add large islands in the deep desert, to stay more consistent with the descriptions in the books.

 

[Deliverator]

One player might be able to work their culture out over a huge area, because of successive spice blooms that just happened to match up next to each other, while another player might not be able to work out at all, because the next spice bloom happened just 1 tile further away.

If you combine it with allowing harvesters to be built outside your cultural borders (by setting bOutsideBorders to 1) then that could get around that problem.

JCultureControl seems to be broken in our mod anyway. I'm trying these settings:

<bSpreadCultureControl>1</bSpreadCultureControl>
<iCultureBorderRange>1</iCultureBorderRange>

but the harvester is not creating a ring of culture - unless I'm mis-understanding how the modcomp works.

Edit: I was misunderstanding. There's a couple more tags.

<iCultureControlStrength>1</iCultureControlStrength>
<iCultureControlCenterTileBonus>2</iCultureControlCenterTileBonus>

It's worth a try at least.

 

[Ahriman]

I was referring to reality according to the Dune books, not necessarily real geography.

I would argue that this isn't islands off in the deep desert, these are islands separated from the other islands by some desert waste basins. They're still real enough to be books: the whole dangerous straits of sand between rocky outcrops is *very* true to the books, like the desert basins that Paul and Jessica cross when they're trying to find the Fremen; and the lack of this sort of thing is part of what I like about the Duneapelago and dislike about the Arrakis script. And they work, gameplay wise.

Attached is a sketch of the sort of thing I had in mind.

 

[Ahriman]

If you combine it with allowing harvesters to be built outside your cultural borders (by setting bOutsideBorders to 1) then that could get around that problem.

Not really, the AI still won't build outside borders, and the AI still won't attack/defend contest the big spice fields.

 

[koma13]

Ahriman, I think you have to open up your mind.
You can see this topic is coming up on a regular basis, indicating flaws in the current spice model.
Let us at least try it, having coexisting concepts have a long tradition in our dune mod.

 

[Deliverator]

the whole dangerous straits of sand between rocky outcrops is *very* true to the books, like the desert basins that Paul and Jessica cross when they're trying to find the Fremen;

I do agree on this point.

I'm going to run some autoplays using Arrakis.py with the following settings on the harvester improvement:

<bOutsideBorders>1</bOutsideBorders>
<bSpreadCultureControl>1</bSpreadCultureControl>
<iCultureBorderRange>1</iCultureBorderRange>
<iCultureControlStrength>1</iCultureControlStrength>
<iCultureControlCenterTileBonus>2</iCultureControlCenterTileBonus>

I'm interested to see what happens.

 

[Deliverator]

These screenshots are result of running 200 turns autoplay with the above settings. It definitely gets the AI out working the deeper desert more.

The key would be getting a worker to seek out spice and build that first harvester even if it's not in your cultural borders. Perhaps we could tweak the AI for that. Give the worker a spice hunter AI mode or something.

I encourage others to try these settings out. More research and less speculation.

 

[Ahriman]

I still have many concerns. For example, I worry that:
a) the culture will stay there even once the harvester disappears, which will look and feel weird
b) the AI will not defend its spice tiles, or attack yours. It would be easy for a human to start a war and pillage off all their harvesters with only a handful of thopters (removing their culture) and then building your own
c) it will be exploitable by the human; the human player can build a huge worker army and go around the whole map getting all the spice, leading to a truly massive economy. The AI won't understand that a huge worker army would be valuable, and so won't build one.
d) it will be very hard to balance the spice income. Either you will be able to run too big an economy with all the spice, or you won't be able to run a big enough one without it.
e) it would allow an arrakis paradise user to still generate a huge spice income. Arrakis paradise penalizes spice income by providing big areas around cities where spice can't spawn, but if you're free to mine all the spice in the deep desert away from your cities then this isn't much of a penalty.

 

[Deliverator]

Your concerns are valid, but it seems that even if we do get a version of Arrakis.py that has more islands, there could still be the issue of spice and spice harvesting not being important enough. It feels like the status quo is not ideal.

a) the culture will stay there even once the harvester disappears, which will look and feel weird

We could add code so that the culture decays once the harvesters disappear perhaps.

b) the AI will not defend its spice tiles, or attack yours. It would be easy for a human to start a war and pillage off all their harvesters with only a handful of thopters (removing their culture) and then building your own

They must be more we can do to encourage the AI to defend/pillage spice harvesters.

c) it will be exploitable by the human; the human player can build a huge worker army and go around the whole map getting all the spice, leading to a truly massive economy. The AI won't understand that a huge worker army would be valuable, and so won't build one.
d) it will be very hard to balance the spice income. Either you will be able to run too big an economy with all the spice, or you won't be able to run a big enough one without it.

We'd need to restrict this ability. Perhaps harvesters built outside cultural borders must still be built within one or two tiles of your current cultural border. We could also copy the city maintenance mechanic and have harvester maintenance that increases with distance from the nearest city or the capital. Part of the Arrakis Spice civic could be reduced harvester maintenance costs.

e) it would allow an arrakis paradise user to still generate a huge spice income. Arrakis paradise penalizes spice income by providing big areas around cities where spice can't spawn, but if you're free to mine all the spice in the deep desert away from your cities then this isn't much of a penalty.

With Arrakis Paradise we can take a more brutal approach to reduce income from spice.

I'm not saying this is the whole solution, but since the problem revolves around cultural borders not extended enough over desert it seems to me that having harvesters create little patches of culture is one of the very few alternatives, and preferable to lots of single rock cities.

I really think we're not going to crack this problem without a few choice AI adjustments, but it is such an important aspect of the game that it is surely worth some custom code.